Comparison of clinical outcomes between hamstring tendon autografts and hybrid grafts in ACL reconstruction: a systematic review and meta-analysis Authors: Lei Fan, Lei Zhang, Jiexi Tang, Zhe Xu, Weili Fu Published: 2024 ABSTRACT BackgroundHamstring tendon (HT) autografts have become a popular choice for anterior cruciate ligament (ACL) reconstruction. However, small-diameter grafts are inevitably encountered during surgery, which have poor biomechanical properties. Hybrid grafts (HGs) using an allograft combined with small diameter HT are gaining interest from surgeons.HypothesisThere would be no difference between the HT autograft and HG in terms of failure, knee stability, and patient-reported outcomes.Study designSystematic review and meta-analysis; Level of evidence, 4.MethodsThe PubMed, Embase, web of science and Cochrane databases were systematically searched from their inception until July 1, 2022. Clinical trials that compared HG and HT autografts were included. The quality of the included studies was assessed with the Cochrane Collaboration’s risk of bias tool and the modified Newcastle-Ottawa Scale. Extracted data were pooled with fixed or random effects depending on the detected heterogeneity.ResultsA total of 14 eligible studies involving 1411 patients (HT: 863; HG: 548) were included in the quantitative meta-analysis. The mean age of the patients involved ranged from 14.6 to 40.4 years. Compared to patients who received HT autografts, patients receiving HGs had similar postoperative failure rate (OR, 0.99; P = 0.97; I2 = 41%), side-to-side difference (MD, -0.16; P = 0.13; I2 = 41%), Subjective IKDC (MD, 0.51; P = 0.58; I2 = 69%), Lysholm (MD, 2.79; P = 0.1; I2 = 79%), Tegner (MD, -0.88; P = 0.56; I2 = 0%). When the available data for failure rate were analyzed by the dose of irradiation, patient age, and mean diameter of the HT, the results of subgroup analyses did not change substantially.ConclusionThis review found no significant differences in failure rates, knee stability, or patient-reported outcomes between autologous HT and HG in ACLR. Surgeons should prioritize autografts of adequate size through optimized techniques and consider hybrid grafts as a last resort, considering the risks associated with allografts. Introduction The rupture of the anterior cruciate ligament (ACL) is one of the most common orthopaedic sports injuries, with approximately 250,000 ACL tears occurring annually in the United States [1]. Delayed surgery increases the risk of meniscal tears and for patients with acute ACL ruptures, especially those engaged in activities involving twisting or rotational movements, early surgical intervention is strongly recommended [2, 3]. ACL reconstruction (ACLR) is a well-established procedure aimed at restoring knee stability, which helps prevent secondary injuries to the meniscus and cartilage, and ultimately reduces the risk of developing osteoarthritis. The choice of graft is critically important as it significantly impacts both surgical outcomes and postoperative rehabilitation [4]. When selecting a graft for ACL reconstruction, surgeons must consider joint stability, functionality, and potential postoperative complications [5]. Ensuring better postoperative knee function scores, such as the Lysholm knee score, Mothadi-QoL, and the Modified Cincinnati Rating Scale, is essential for optimizing patient outcomes [6]. Among the various graft options available for ACLR, the use of autologous hamstring tendon (HT) remains the most prevalent, primarily due to its favorable clinical outcomes and lower incidence of donor-site morbidity compared to bone–patellar tendon–bone (BPTB) grafts. While patellar tendon grafts can achieve satisfactory knee-abduction moment (KAM) values within the physiological range postoperatively, their use in ACL reconstruction may result in permanent loss of bone mineral content due to reduced patellar tendon stiffness [7, 8]. Furthermore, HT grafts are associated with a reduced risk of mid- and long-term osteoarthritis development [9, 10]. However, one of the challenges associated with autologous HT is the variability in graft size, particularly when intraoperative harvesting results in smaller diameter grafts due to individual differences or iatrogenic injury during the procedure. Although the ideal graft diameter has not been definitively established, current consensus suggests that grafts with a diameter less than 8 mm are associated with an increased risk of failure, as biomechanical strength is generally proportional to the graft’s diameter [11, 12]. Research has shown that increasing the HT graft diameter by just 0.5 mm reduces the likelihood of revision surgery by approximately 0.82 to 0.83 times [13, 14]. Given the unpredictability of tendon size before surgery, surgeons are often faced with the need to augment undersized grafts using additional tissue to achieve the desired diameter. In response to this challenge, hybrid grafts (HGs), which typically combine autologous HT with soft-tissue allografts to customize the graft size, have gained increasing attention among orthopedic surgeons. HGs offer a theoretical advantage by preserving the benefits of autografts while potentially reducing the failure rate of ACLR by ensuring a more suitable graft size. Harvesting the HT does not generally lead to clinical disability or decreased athletic performance. Nevertheless, for certain patients, preserving a fully functional HT may offer additional benefits. When the use of a HG is planned in advance, surgeons can mitigate potential morbidity by harvesting only one of the HTs, leaving the other intact [15]. Despite these advantages, the use of HGs also presents notable disadvantages. A significant concern is the treatment of most allografts with radiation, which can weaken the graft’s biomechanical properties and increase the risk of graft failure [16, 17]. Additionally, clinical studies comparing the outcomes of HGs and autologous HTs have yielded conflicting results [18–21]. These discrepancies may be attributed to variations in allograft processing methods and differences in the patient populations included in the studies [18, 21–23]. Limited evidence exists to suggest that HGs are superior to autologous HTs in ACLR, and most systematic reviews and meta-analyses comparing the two grafts have predominantly focused on graft failure rates [24, 25]. Given the current body of literature, the clinical outcomes of HGs versus HTs in ACLR remain a topic of ongoing debate. Factors such as patient-reported outcomes, knee stability, and the long-term success of the grafts are crucial considerations in evaluating the efficacy of these approaches. In light of these concerns, this study aims to systematically review and evaluate the clinical outcomes of HGs versus autologous HTs in ACLR, specifically focusing on failure rates, patient-reported outcomes, and knee stability. Based on both our clinical experience and the current evidence, we hypothesize that there will be no significant difference in the overall outcomes between these two graft types. Methods We performed and conducted this review according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The protocol was registered at PROSPERO before starting this review (CRD42022345704). Literature search The PubMed, Embase, web of science and Cochrane databases were searched for any relevant studies published before July 1, 2022. Search terms included “anterior cruciate ligament”, “ACL”, “anterior cruciate ligament reconstruction”, “hamstring”, “hamstring tendon”, “hamstring tendon autograft” and “hybrid”. Two reviewers independently screened the studies for the eligibility of identified studies. Any discrepancies were resolved by consensus between the reviewers. The senior author (W.F.) was consulted in case of persistent disagreement. The gray literature and databases of unpublished studies were also examined, as well as the reference lists of all included studies were also screened for potentially eligible studies. Inclusion criteria for eligible studies in this review were as follows: (1) all comparative studies (randomized controlled trials [RCTs], observational studies) of human patients undergoing primary ACLR; (2) a direct comparison of hamstring autografts alone and autograft-allograft HGs; (3) at least 1 validated clinical outcome was reported (failure, patient-reported outcomes, knee stability, etc.); (4) studies with a minimum 1-year follow-up. Exclusion criteria were as follows: (1) revision ACLR; (2) primary ACLR with allograft or artificial ligaments; (3) articles written in a language other than English; (4) non-clinical studies (e.g., cadaveric or animal model); (5) case reports, case series, commentaries, and reviews. Study with smallest cohort or shortest follow-up (different studies that focus on the same group of patients). Data extraction After identifying all eligible studies, the same 2 reviewers created a spreadsheet to independently extract relevant information: (1) publication information (first author, year of publication, study design, length of follow-up); (2) patient information (sample size, age, sex); (3) surgical procedure (graft type, method of fixation and graft diameter); (4) surgical outcomes (failure, patient-reported outcomes, knee stability, etc.). Patient-reported outcomes include Subjective International Knee Documentation Committee (IKDC) Score, Lysholm Score and Tegner Score. Knee stability measures include the side-to-side difference in translation, Lachman test and pivot-shift test. Failure of ACLR was reported in all 14 studies. Graft failure was defined based on postoperative MRI, second-look arthroscopy, physical examination, and/or patient-reported outcomes. In 4 studies [18, 22, 26, 27], failure was defined as revision ACLR or complete graft rupture on MRI or at the time of arthroscopy. In 2 studies [28, 29], the criteria for defining failure were supplemented with patient-reported outcomes and physical examination. In 1 study [30], failure was defined as a failure in the completion of the ligamentization process, leading to an atonic, disorganized, and non-viable graft. Seven studies [19–21, 23, 31–33] did not clearly define the failure. Any values expressed as median, interquartile range, or 95% CI were converted to mean ± Standard Deviation (SD) with mathematical formulas and Review Manager 5.3 software. If necessary, the corresponding author of the included studies will be contacted to obtain the original data. P values less than 0.05 were considered to be statistically significant. Any disagreements on data interpretation were resolved by discussion or consultation with the senior author (W.F.). Risk of bias assessment The same 2 reviewers independently assessed the quality of included studies. The Cochrane Collaboration’s risk of bias tool was used to assess the risk of bias of RCTs [34], including random sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective reporting and other bias. The modified Newcastle-Ottawa Scale (NOS) was used to assess the risk of bias in all the nonrandomized cohort studies [35], including group selection, group comparability, and ascertainment of the outcome of interest. A star indicates that the study met the requirements of the item. A maximum of 9 stars could be given to each study. Studies with scores of ≥ 7, 5 to 7, 3 to 5, and 0 to 2 were considered good, fair, poor-fair, and poor quality, respectively. The modified Coleman methodology score (MCMS) was also used to assess the methodological quality of each study [36]. Each author used the tool to assign the score for each study out of 100. The maximum score of 100 indicates a study that avoids confounding factors and bias. Scores ranging from 85 to 100 are excellent, 70 to 84 are good, 55 to 69 are fair, and less than 55 are poor. Data synthesis The Review Manager 5.3 (The Cochrane Collaboration, 2013) was used to analyze all extracted data. Dichotomous outcomes were reported as odds ratios (OR) with 95% confidence intervals (CI), while continuous outcomes were analyzed as mean differences (MD) with 95% CI. Heterogeneity between studies was quantified by I2. I2 > 50% indicated high heterogeneity; the random-effects model was used. Otherwise, the fixed-effects model was applied. When data for that outcome could be pooled from more than 10 studies, funnel plots were used to assess potential publication bias. Sensitivity analysis was used to assess the impact of individual included studies on the pooled results. Results Search results The search strategy generated a total of 1541 articles. After removing the 299 duplicate studies, 1199 of the remaining studies were excluded based on their titles and abstracts. A total of 43 articles were recruited for full-text screening. Eventually, 14 articles were included in this systematic review and meta-analysis:1 RCT [32] and 13 observational studies [18–23, 26–31, 33]. Figure 1 shows The PRISMA diagram of the article search and selection process. Fig. 1Flow diagram Flow diagram Study characteristics A total of 1411 patients were included in this meta-analysis, with mean age from 14.6 to 40.4 years. Of whom, 863 received HTs and 548 received HGs. The publication dates of the included studies ranged from 2015 to 2021, and the mean follow-up period ranged from 19.4 months to 70.8 months. The allograft tissues comprised of the hybrid graft included HT, peroneus longus, tibialis anterior and tibialis posterior. A total 12 studies [18, 20–23, 26–30, 32, 33] in which allografts were treated with irradiation, the irradiation dose ranges from 1.2 Mrad to 2.5 Mrad. Nine studies reported that the diameter of the HGs was significantly greater than that of the HTs [18, 20–22, 26–28, 30, 31] and 2 studies did not report the diameter of grafts [19, 32]. All studies used the same tibial and femoral fixation method for both the HT autografts and HGs. The most common method of fixing femoral-side grafts was suspensory fixation, while the tibial end is mostly fixed with interference screws. All study baseline characteristics are summarized in Tables 1 and 2. Table 1Basic characteristics of included studies in the analysisFirst author, yearStudy designLOEDuration of studySample size, n (male)Patient age, y (Mean ± SD)Follow up, monOutcomesHAHGHAHGHAHGBurrus et al. 2015RCS32007–201229 (10)29 (10)27.2 ± 9.626.6 ± 10.348 ± 15.244.4 ± 16.9IKDC, Lysholm, failure, reoperationDarnley et al. 2016Cohort study32008–201127 (17)27 (17)20.9 ± 7.020.9 ± 7.02424IKDC, KOOS, Marx activity score, failureHeyworth et al. 2021RCS32003–201764 (5)47 (17)16.1 ± 1.515.8 ± 1.927.6 ± 20.432.4 ± 27.6failure, time to return to sports, radiographyJacobs et al. 2016Cohort study32010–201546 (20)42 (24)14.6 ± 1.715.0 ± 1.846.7 ± 31.931.2 ± 13.8failureKraeutler et al. 2018RCS32001–2013836528.6 ± 9.940.4 ± 10.3*56.4 ± 22.842.0 ± 10.8IKDC, KOOS, SANE, SF-12,failures, reoperationsLeo et al. 2016RCS32010–201371 (46)24 (16)26.9 ± 3.328.1 ± 5.326.926.9Operating time, IKDC, UCLA, ACL-QL, VAS, failure, reoperationLi et al. 2015RCT22005–200832 (17)31 (18)29.8 ± 7.931.6 ± 8.269.6 ± 10.870.8 ± 7.2Operating time, IKDC, Tegner, Lysholm, pivot shift, Lachman, KT-1000Pennock et al. 2016Cohort study32003–201420 (7)20 (8)15.9 ± 1.715.4 ± 1.83636failure, Lysholm, Tegner, SANE, satisfactionPerkins et al. 2019RCS32012–2016289 (139)65 (18)NR16.0 ± 1.526 (6–56) bfailureRao et al. 2021RCS32005–201580 (55)59 (12)22.1 (17.6, 38.2) a23.2 (17, 38.8) a≥ 2 yearsrevison, reoperation, IKDC, Tegner, NPRS, SANEWang et al. 2018RCS32013–201429 (10)28 (9)32.6 ± 10.332.8 ± 9.940.6 ± 3.540.6 ± 3.4IKDC, Tegner, Lysholm, pivot shift,Lachman, KT-1000, failure, reoperationXu et al. 2017RCS32014–201731 (16)37 (27)32.8 ± 8.933.9 ± 8.428.4 ± 2.128.3 ± 2.8IKDC, Tegner, Lysholm, KT-1000, graft continuity, graft tension, graftsynovial coverage, K-L grade, failure, reoperationXu et al. 2018RCS32013–201434 (21)42 (31)35.32 ± 9.8735.93 ± 9.6527.02 ± 1.8826.84 ± 1.83Lysholm, Tegner, IKDC, KT-1000Zheng et al. 2019RCS32013–201828 (23)32 (20)30.1 ± 9.629.0 ± 8.421.1 + 7.319.4 + 6.2Lysholm, IKDC, Tegner, KT-1000, second-look arthroscopic findingsData were presented as median, range a or median (interquartile range) bACL, anterior cruciate ligament; ACL-QL, ACL-Quality of Life; HA, hamstring autograft; HG, hybrid graft; IKDC, International Knee Documentation Committee; K-L, Kellgren-Lawrence; KOOS, Knee Injury and Osteoarthritis Outcome Score; LOE, level of evidence; NR, not reported; NPRS, Numeric Pain Rating Scale; RCS, Retrospective comparative study; SANE, Single Assessment Numeric Evaluation; SF-12, Short Form 12; UCLA, University of California, Los Angeles; VAS, visual analog scale Table 1Basic characteristics of included studies in the analysisFirst author, yearStudy designLOEDuration of studySample size, n (male)Patient age, y (Mean ± SD)Follow up, monOutcomesHAHGHAHGHAHGBurrus et al. 2015RCS32007–201229 (10)29 (10)27.2 ± 9.626.6 ± 10.348 ± 15.244.4 ± 16.9IKDC, Lysholm, failure, reoperationDarnley et al. 2016Cohort study32008–201127 (17)27 (17)20.9 ± 7.020.9 ± 7.02424IKDC, KOOS, Marx activity score, failureHeyworth et al. 2021RCS32003–201764 (5)47 (17)16.1 ± 1.515.8 ± 1.927.6 ± 20.432.4 ± 27.6failure, time to return to sports, radiographyJacobs et al. 2016Cohort study32010–201546 (20)42 (24)14.6 ± 1.715.0 ± 1.846.7 ± 31.931.2 ± 13.8failureKraeutler et al. 2018RCS32001–2013836528.6 ± 9.940.4 ± 10.3*56.4 ± 22.842.0 ± 10.8IKDC, KOOS, SANE, SF-12,failures, reoperationsLeo et al. 2016RCS32010–201371 (46)24 (16)26.9 ± 3.328.1 ± 5.326.926.9Operating time, IKDC, UCLA, ACL-QL, VAS, failure, reoperationLi et al. 2015RCT22005–200832 (17)31 (18)29.8 ± 7.931.6 ± 8.269.6 ± 10.870.8 ± 7.2Operating time, IKDC, Tegner, Lysholm, pivot shift, Lachman, KT-1000Pennock et al. 2016Cohort study32003–201420 (7)20 (8)15.9 ± 1.715.4 ± 1.83636failure, Lysholm, Tegner, SANE, satisfactionPerkins et al. 2019RCS32012–2016289 (139)65 (18)NR16.0 ± 1.526 (6–56) bfailureRao et al. 2021RCS32005–201580 (55)59 (12)22.1 (17.6, 38.2) a23.2 (17, 38.8) a≥ 2 yearsrevison, reoperation, IKDC, Tegner, NPRS, SANEWang et al. 2018RCS32013–201429 (10)28 (9)32.6 ± 10.332.8 ± 9.940.6 ± 3.540.6 ± 3.4IKDC, Tegner, Lysholm, pivot shift,Lachman, KT-1000, failure, reoperationXu et al. 2017RCS32014–201731 (16)37 (27)32.8 ± 8.933.9 ± 8.428.4 ± 2.128.3 ± 2.8IKDC, Tegner, Lysholm, KT-1000, graft continuity, graft tension, graftsynovial coverage, K-L grade, failure, reoperationXu et al. 2018RCS32013–201434 (21)42 (31)35.32 ± 9.8735.93 ± 9.6527.02 ± 1.8826.84 ± 1.83Lysholm, Tegner, IKDC, KT-1000Zheng et al. 2019RCS32013–201828 (23)32 (20)30.1 ± 9.629.0 ± 8.421.1 + 7.319.4 + 6.2Lysholm, IKDC, Tegner, KT-1000, second-look arthroscopic findingsData were presented as median, range a or median (interquartile range) bACL, anterior cruciate ligament; ACL-QL, ACL-Quality of Life; HA, hamstring autograft; HG, hybrid graft; IKDC, International Knee Documentation Committee; K-L, Kellgren-Lawrence; KOOS, Knee Injury and Osteoarthritis Outcome Score; LOE, level of evidence; NR, not reported; NPRS, Numeric Pain Rating Scale; RCS, Retrospective comparative study; SANE, Single Assessment Numeric Evaluation; SF-12, Short Form 12; UCLA, University of California, Los Angeles; VAS, visual analog scale Basic characteristics of included studies in the analysis IKDC, KOOS, SANE, SF-12, failures, reoperations Operating time, IKDC, Tegner, Lysholm, pivot shift, Lachman, KT- 1000 IKDC, Tegner, Lysholm, pivot shift, Lachman, KT-1000, failure, reoperation IKDC, Tegner, Lysholm, KT-1000, graft continuity, graft tension, graft synovial coverage, K-L grade, failure, reoperation Data were presented as median, range a or median (interquartile range) b ACL, anterior cruciate ligament; ACL-QL, ACL-Quality of Life; HA, hamstring autograft; HG, hybrid graft; IKDC, International Knee Documentation Committee; K-L, Kellgren-Lawrence; KOOS, Knee Injury and Osteoarthritis Outcome Score; LOE, level of evidence; NR, not reported; NPRS, Numeric Pain Rating Scale; RCS, Retrospective comparative study; SANE, Single Assessment Numeric Evaluation; SF-12, Short Form 12; UCLA, University of California, Los Angeles; VAS, visual analog scale Table 2Details of the grafts and surgeryFirst author, yearIrradiation, MradIndication for hybridGraft diameter, mmGraft typeFixationHAHGHAHGFemoralTibialBurrus et al. 20151.2–1.9< 7.5 mm7.8 ± 0.79.2 ± 0.9*HTHT + ST allo or TA alloSuspensory or cross-pininterference screwDarnley et al. 20161.2–1.9< 8 mm8.4 (7–10)9.5 (8–10) *HTST or HT or GT + ST alloSuspensoryinterference screw + staple/ screw and washerHeyworth et al. 2021NR< 8 mm7.1 ± 0.29.1 ± 0.9*HTHT + TA allo or TP allo or ST or GTSuspensoryinterference screwJacobs et al. 20161.2–1.9NR7.8 ± 1.29.9 ± 0.8*HTHT + ST alloSuspensoryinterference screwKraeutler et al. 2018NonirradiatedNR7.5 ± 0.59.1 ± 0.5*HTST + TP allo or PL alloSuspensoryinterference screw /spiked washer and screwLeo et al. 2016Nonirradiated< 8 mmNRNRHTHT + ST allo or TA allocross-pininterference screwLi et al. 20152.5NRNRNRSTST + TA alloSuspensoryinterference screwPennock et al. 20160.9–1.3< 7 mm6.48.9*HTHT + ST allo or TA alloSuspensory + TightRopeinterference screwPerkins et al. 20190.9–1.3< 7.5 mm8.6 (7.5–10)9.2 (8–10) *STST + ST allo or TP alloSuspensoryinterference screwRao et al. 2021Low dose gamma irradiation< 8 mm9 (9, 9.5)10 (9, 11)HTHT + HT alloSuspensory (99.3%)interference screw (35.3%) or spiked washer/screw (56.1%)Wang et al. 20182.5< 8 mm8.89.1HTHT + TA alloSuspensoryinterference screwXu et al. 20172.5NR8.5 ± 0.79 ± 0.5*HTHT + TA alloSuspensoryinterference screwXu et al. 20182.5< 8 mm8.57 ± 0.489.1 ± 0.52STGSTG + TA alloSuspensoryinterference screwZheng et al. 20192.5< 8 mm8.5 ± 0.59 ± 0.6*HTHT + TA alloSuspensory + TightRopeinterference screw*Statistically significant difference between the 2 groupsData are presented as mean ± SD, or mean (range), or median (IQR)allo, allograft; GT, gracilis tendon; HA: hamstring autograft; HG: hybrid graft; HT, hamstring tendon; NR, not reported; PL, Peroneus Longus; ST, semitendinosus tendon; TA, tibialis anterior; TP, tibialis posterior Table 2Details of the grafts and surgeryFirst author, yearIrradiation, MradIndication for hybridGraft diameter, mmGraft typeFixationHAHGHAHGFemoralTibialBurrus et al. 20151.2–1.9< 7.5 mm7.8 ± 0.79.2 ± 0.9*HTHT + ST allo or TA alloSuspensory or cross-pininterference screwDarnley et al. 20161.2–1.9< 8 mm8.4 (7–10)9.5 (8–10) *HTST or HT or GT + ST alloSuspensoryinterference screw + staple/ screw and washerHeyworth et al. 2021NR< 8 mm7.1 ± 0.29.1 ± 0.9*HTHT + TA allo or TP allo or ST or GTSuspensoryinterference screwJacobs et al. 20161.2–1.9NR7.8 ± 1.29.9 ± 0.8*HTHT + ST alloSuspensoryinterference screwKraeutler et al. 2018NonirradiatedNR7.5 ± 0.59.1 ± 0.5*HTST + TP allo or PL alloSuspensoryinterference screw /spiked washer and screwLeo et al. 2016Nonirradiated< 8 mmNRNRHTHT + ST allo or TA allocross-pininterference screwLi et al. 20152.5NRNRNRSTST + TA alloSuspensoryinterference screwPennock et al. 20160.9–1.3< 7 mm6.48.9*HTHT + ST allo or TA alloSuspensory + TightRopeinterference screwPerkins et al. 20190.9–1.3< 7.5 mm8.6 (7.5–10)9.2 (8–10) *STST + ST allo or TP alloSuspensoryinterference screwRao et al. 2021Low dose gamma irradiation< 8 mm9 (9, 9.5)10 (9, 11)HTHT + HT alloSuspensory (99.3%)interference screw (35.3%) or spiked washer/screw (56.1%)Wang et al. 20182.5< 8 mm8.89.1HTHT + TA alloSuspensoryinterference screwXu et al. 20172.5NR8.5 ± 0.79 ± 0.5*HTHT + TA alloSuspensoryinterference screwXu et al. 20182.5< 8 mm8.57 ± 0.489.1 ± 0.52STGSTG + TA alloSuspensoryinterference screwZheng et al. 20192.5< 8 mm8.5 ± 0.59 ± 0.6*HTHT + TA alloSuspensory + TightRopeinterference screw*Statistically significant difference between the 2 groupsData are presented as mean ± SD, or mean (range), or median (IQR)allo, allograft; GT, gracilis tendon; HA: hamstring autograft; HG: hybrid graft; HT, hamstring tendon; NR, not reported; PL, Peroneus Longus; ST, semitendinosus tendon; TA, tibialis anterior; TP, tibialis posterior Details of the grafts and surgery *Statistically significant difference between the 2 groups Data are presented as mean ± SD, or mean (range), or median (IQR) allo, allograft; GT, gracilis tendon; HA: hamstring autograft; HG: hybrid graft; HT, hamstring tendon; NR, not reported; PL, Peroneus Longus; ST, semitendinosus tendon; TA, tibialis anterior; TP, tibialis posterior Assessment of risks of bias Only 1 study was a randomized controlled trial, and it did not clearly report the procedure of randomization and concealment of allocation and was rated as unclear risk in selection bias. Due to differences in graft diameter, it is not possible to blind the surgeon performing the procedure and was rated as unclear risk in performance bias. The risk bias assessment for all other items was rated as low risk. The NOS of non-RCTs ranged from 7 to 9 stars. Most studies scored 8 stars, indicating good quality and low risk of bias (Table 3). Table 3Newcastle-Ottawa Scale assessing the quality of studies included in the systematic reviewFirst author, yearSelectionComparabilityOutcomeRepresentativeness of treated cohortSelection of comparative cohortAscertainment of treated cohort recordsThe outcome of interest was not present at startControls for ageControls for any additional factorAssessment of outcomeLong enough follow-upAdequacy of follow-upTotal quality scoreBurrus et al. 2015★★★★★★★7Darnley et al. 2016★★★★★★★★8Heyworth et al. 2021★★★★★★★★8Jacobs et al. 2016★★★★★★★★8Kraeutler et al. 2018★★★★★★★★8Leo et al. 2016★★★★★★★★8Pennock et al. 2016★★★★★★★7Perkins et al. 2019★★★★★★★★8Rao et al. 2021★★★★★★★★8Wang et al. 2018★★★★★★★★★9Xu et al. 2017★★★★★★★7Xu et al. 2018★★★★★★★★8Zheng et al. 2019★★★★★★★7 Table 3Newcastle-Ottawa Scale assessing the quality of studies included in the systematic reviewFirst author, yearSelectionComparabilityOutcomeRepresentativeness of treated cohortSelection of comparative cohortAscertainment of treated cohort recordsThe outcome of interest was not present at startControls for ageControls for any additional factorAssessment of outcomeLong enough follow-upAdequacy of follow-upTotal quality scoreBurrus et al. 2015★★★★★★★7Darnley et al. 2016★★★★★★★★8Heyworth et al. 2021★★★★★★★★8Jacobs et al. 2016★★★★★★★★8Kraeutler et al. 2018★★★★★★★★8Leo et al. 2016★★★★★★★★8Pennock et al. 2016★★★★★★★7Perkins et al. 2019★★★★★★★★8Rao et al. 2021★★★★★★★★8Wang et al. 2018★★★★★★★★★9Xu et al. 2017★★★★★★★7Xu et al. 2018★★★★★★★★8Zheng et al. 2019★★★★★★★7 Newcastle-Ottawa Scale assessing the quality of studies included in the systematic review Methodological quality assessment The MCMS of the included studies were shown in Table 4. Most included studies showed clear methodological flaws and did not achieve excellent scores. The mean MCMS was 67 (range, 56–85). The primary reasons for the low scores were sample size, patient selection process, and study type. Table 4Modified coleman methodology score assessing the quality of studies included in the systematic reviewFirst author, yearCriteria part ACriteria part BTotalStudy sizeMean duration of follow-upSurgical approachType of studyDiagnostic certaintyDescription of surgical procedureDescription of rehabilitation proceduresOutcome criteriaProcedure for assessing outcomesSubject selection processTotal (part A)Total (part B)Total scoreBurrus et al. 2015777051058125412566Darnley et al. 20167471051058125482573Heyworth et al. 2021104705558125362561Jacobs et al. 201677101051057125542478Kraeutler et al. 2018107705555125392261Leo et al. 2016741005558155362864Li et al. 2015710101551056125622385Pennock et al. 2016441010550895382260Perkins et al. 2019104705105595411960Rao et al. 2021104705058125312556Wang et al. 201877100510510155443074Xu et al. 20177410051058125412566Xu et al. 20187410051058125412566Zheng et al. 20197410055510125362763 Table 4Modified coleman methodology score assessing the quality of studies included in the systematic reviewFirst author, yearCriteria part ACriteria part BTotalStudy sizeMean duration of follow-upSurgical approachType of studyDiagnostic certaintyDescription of surgical procedureDescription of rehabilitation proceduresOutcome criteriaProcedure for assessing outcomesSubject selection processTotal (part A)Total (part B)Total scoreBurrus et al. 2015777051058125412566Darnley et al. 20167471051058125482573Heyworth et al. 2021104705558125362561Jacobs et al. 201677101051057125542478Kraeutler et al. 2018107705555125392261Leo et al. 2016741005558155362864Li et al. 2015710101551056125622385Pennock et al. 2016441010550895382260Perkins et al. 2019104705105595411960Rao et al. 2021104705058125312556Wang et al. 201877100510510155443074Xu et al. 20177410051058125412566Xu et al. 20187410051058125412566Zheng et al. 20197410055510125362763 Modified coleman methodology score assessing the quality of studies included in the systematic review Outcome synthesis Failure rate All 14 included studies with 1411 patients reported graft failure rate. In the HT group, the graft failure rate ranged from 0 to 28.3%; in the HG group, the graft failure rate ranged from 0 to 30%. HT group showed an overall failure rate of 11.2% (97/863) and the HG group showed an overall failure rate of 10% (55/548). No significant difference in failure rate was noted between the 2 groups (OR, 0.99; 95% CI, 0.69–1.43; P = 0.97; I2 = 41%) (Fig. 2). Fig. 2Failure rate Failure rate Failure rate Side-to-side difference A KT-1000 arthrometer was performed to measure the side-to-side difference of knee anterior translation. A total of 5 studies with 324 patients reported the postoperative side-to-side difference. The pooled results showed no significant difference between the 2 groups (MD, -0.16; 95% CI, -0.36-0.05; P = 0.13; I2 = 41%) (Fig. 3). Fig. 3Side-to-side difference Side-to-side difference Lachman Test (Grade 0) and Pivot-Shift Test (Grade 0) Only two studies involving 120 patients reported the Lachman test and the Pivot-Shift test. No significant difference was observed between HA and HG groups in terms of the Grade 0 Lachman test (OR, 2.03; 95% CI, 0.22–18.88; P = 0.53; I2 = 82%) and the Grade 0 Pivot-Shift test (OR, 2.28; 95% CI, 0.27–18.92; P = 0.45; I2 = 75%) (Fig. 4). Fig. 4Lachman test and Pivot-Shift test Lachman test and Pivot-Shift test Patient-Reported Outcomes Subjective IKDC score A total of 10 studies with 818 patients reported the postoperative Subjective IKDC Score. The mean Subjective IKDC Score in the HA group was 78.5, and the mean Subjective IKDC Score in the HG group was 76.9. The pooled results showed no significant difference between the 2 groups (MD, 0.51; 95% CI, -1.31-2.33; P = 0.58; I2 = 69%) (Fig. 5A). The statistical heterogeneity for this outcome was reported to be substantial, indicating the huge diversity of the studies included. Heterogeneity was dramatically reduced after removing the studies by Burrus et al. [26] and wang et al. [29], and pooled analysis of the remaining studies still showed no significant change from the original results (MD, -0.56; 95% CI, -1.56-0.43; P = 0.27; I2 = 8%) (Fig. 5B). Fig. 5A IKDC Score. B IKDC Score (reduced heterogeneity) A IKDC Score. B IKDC Score (reduced heterogeneity) Lysholm score A total of 7 studies with 422 patients reported the postoperative Lysholm Score. The mean Lysholm Score in the HA group was 89.7, and the mean Lysholm Score in the HG group was 86.6. The pooled results showed no significant difference between the 2 groups (MD, 2.79; 95% CI, -0.55-6.13; P = 0.1; I2 = 79%) (Fig. 6A). Heterogeneity was also dramatically reduced after removing the studies by Burrus et al. [26] and Wang et al. [29], and pooled analysis of the remaining studies showed no significant change from the original results (MD, -0.06; 95% CI, -1.67-1.56; P = 0.95; I2 = 33%) (Fig. 6B). Fig. 6A Lysholm Score. B Lysholm Score (reduced heterogeneity) A Lysholm Score. B Lysholm Score (reduced heterogeneity) Tegner score A total of 7 studies with 503 patients reported the postoperative Tegner Score. All 7 studies showed no significant difference in this score between the two groups. The mean Tegner Score in the HA group was 6.1, and the mean Tegner Score in the HG group was 6.2. The pooled results showed no significant difference between the 2 groups (MD, -0.88; 95% CI, -0.35-0.19; P = 0.56; I2 = 0%) (Fig. 7). Fig. 7Tegner Score Tegner Score Subgroups analysis Some articles have reported that radiation to allografts may affect the biomechanical properties of the grafts and that this effect is dose-dependent [37, 38]. Besides, adolescents have been identified as an independent risk factor for re-rupture following primary ACLR. And biologic responses to allograft tissue and activity levels differed significantly between adolescents and adults. Variations in autograft diameter among studies also contributed to some of the heterogeneity. Therefore, we performed subgroup analyses of the primary outcome - failure rate - based on the irradiation dose, patient age, and mean diameter of the HT. None of the results changed substantially, proving that our statistical results are robust. Table 5 summarized the results of the failure rate subgroup analyses. Table 5Subgroup analyses of failure rateVariableNo. of studiesFailure rateOdds ratios (95% CI)P valueI2, %HAHGAge<20461/41928/1740.94 [0.34, 2.66]0.9166≥ 201039/58231/5241.08 [0.64, 1.85]0.7728Dose of irradiation, Mrad0214/1543/893.09 [0.87, 10.99]0.080<2.5671/49140/2420.76 [0.33, 1.76]0.5260≥ 2.554/1548/1700.52 [0.15, 1.78]0.30diameter of autografts (mm)<8 mm533/24221/2031.15[0.36–3.70]0.8163≥ 8 mm760/51833/2900.76[0.47–1.23]0.274 Table 5Subgroup analyses of failure rateVariableNo. of studiesFailure rateOdds ratios (95% CI)P valueI2, %HAHGAge<20461/41928/1740.94 [0.34, 2.66]0.9166≥ 201039/58231/5241.08 [0.64, 1.85]0.7728Dose of irradiation, Mrad0214/1543/893.09 [0.87, 10.99]0.080<2.5671/49140/2420.76 [0.33, 1.76]0.5260≥ 2.554/1548/1700.52 [0.15, 1.78]0.30diameter of autografts (mm)<8 mm533/24221/2031.15[0.36–3.70]0.8163≥ 8 mm760/51833/2900.76[0.47–1.23]0.274 Subgroup analyses of failure rate Publication bias A funnel plot was applied to evaluate publication bias. The funnel plots for studies reporting failure rate and subjective IKDC score were symmetric, indicating a low risk of publication bias (Fig. 8). Fig. 8Publication bias Publication bias Discussion This systematic review provided a comprehensive review of clinical outcomes comparing autologous HT and HG in patients undergoing ACLR. The pooled results from 14 studies involving 1411 patients demonstrated similar results in all outcomes in patients undergoing ACLR with a HT compared to a HG. There are no statistically significant differences in failure rate, KT-1000 side-to-side difference, Lachman test, pivot-shift test, subjective IKDC score, Lysholm score, or Tegner score. The findings of the current study are consistent with previously published systematic reviews. To our knowledge, this is the first systematic review and meta-analysis that comprehensively assessed relevant outcomes between HT and HG in ACLR. In evaluating the surgical failure rates, we found no significant difference between the two groups, although the literature indicates a rising trend in failure rates. For example, Perkins’ study reported that hybrid autologous hamstring-allograft tendon grafts had a 2.6-fold higher risk of failure compared to autologous hamstring tendon grafts. Similar findings were reported by Burrus MT, Pennock, and Wang, emphasizing the importance of tendon-bone healing and collagen fiber remodeling for the biological and mechanical performance of the graft [26, 29, 39, 40]. The tendon-bone healing process involves four stages: inflammation, proliferation, matrix synthesis, and remodeling. However, in hybrid autologous-allograft grafts, inconsistencies in cell death, blood flow, and revascularization can impair the healing process [41, 42]. Scheffler SU further noted that ACL grafts undergo necrosis, revascularization, cellular proliferation, and collagen remodeling, with the quality of ligamentization directly influencing the graft’s long-term survival [43]. Yet, the slower revascularization and ligamentization seen in allografts, particularly in hybrid grafts, may weaken mechanical performance [43, 44]. Wang et al. observed differences in graft maturity during secondary arthroscopy, with hybrid grafts showing reduced remodeling and ligamentization, which can compromise long-term stability and functionality [29, 43]. Ge Y found that two years post-surgery, 3.0-T MRI 3D modeling revealed poorer remodeling in allografts within the bone tunnel, likely due to immune reactions to foreign substances [45]. Damaged grafts also lost part of their reconstructive function, which may contribute to graft failure [26]. Synovialization plays a key role in graft healing, with autologous hamstring tendon grafts showing significantly better synovial coverage than allografts, and this degree of synovialization is positively correlated with clinical outcomes [46, 47]. The processing of allografts, such as through dehydration, deproteinization, or irradiation, can further impair their clinical efficacy [48]. Irradiation, in particular, has a dose-dependent negative impact on graft strength, with irradiated allografts performing worse mechanically than non-irradiated ones [49, 50]. The positioning of allografts relative to autografts in hybrid grafts may also influence tendon-bone healing, with slower biological healing seen in allograft tissues [51]. Kaeding CC warned that young, active patients who return to high-intensity activities before the graft is fully healed are at greater risk of failure. Therefore, patients receiving allograft reconstructions should allow for extended recovery times to ensure better graft integration [52]. However, the study by Baz et al. [53]. demonstrated that 3D-assisted surgery positively impacts surgical time, blood loss, fluoroscopy use, and complication rates. The use of pre-shaped or patient-specific implants also contributes to improved efficiency. Similarly, in ACL reconstruction, 3D-assisted surgery could ensure more precise implant placement by comprehensively assessing graft diameter, length, and other relevant characteristics preoperatively and intraoperatively. This may enhance remodeling in allografts within the bone tunnel and improve reconstructive function, though further research is needed to confirm these findings. Graft diameter and occupancy are critical factors in the success of ACL reconstruction, yet simply increasing the diameter of an allograft may not improve function or stability. Xu emphasized that graft purity is likely more important than size during the vascularization and tendon-bone healing processes. Definitions of graft failure vary across studies, with Jacobs defining failure as patient-reported instability that affects daily or athletic activities, combined with pathological laxity during physical exams, or evidence of graft failure from MRI or arthroscopy [54, 55]. Perkins focused on ACL revision patients, identifying failure through complete graft rupture on MRI or surgeon-reported revision surgery [39]. Zheng confirmed ACL re-rupture via physical examination, MRI, and side-to-side joint measurements, while Burrus MT reported seven cases of failure confirmed by MRI and one by arthroscopy [26, 30]. These differing definitions and diagnostic methods could lead to variations in reported failure rates. Aside from graft diameter, other factors influencing failure include associated injuries, surgical techniques, such as fixation methods, tunnel positioning, and the independent or combined effects of anterior positioning of allograft tissue, rehabilitation protocols, and patient-specific characteristics like body weight and height. These confounding factors warrant further exploration, and more high-quality studies are needed to provide definitive conclusions. In the knee stability analysis, we evaluated side-to-side differences, the Lachman test, and the pivot shift test, and found no significant differences between the two groups. Li’s study showed better stability in the hybrid graft group, while Wang’s study favored the autograft group, but neither reached statistical significance [29, 56]. Li further reported that both autograft and hybrid grafts performed better than allografts in KT-1000 measurements, with the hybrid graft showing slightly better anterior-posterior stability, though not statistically significant. Boniello MR’s biomechanical study indicated that increased graft diameter enhances tensile strength and reduces laxity, while Prodromos C found that irradiated grafts had lower stability compared to non-irradiated grafts [57, 58]. Zheng noted that knee laxity differences between the hybrid and allograft groups decreased over time due to similar synovialization levels [30]. Wang’s second-look arthroscopy revealed graft maturity differences, suggesting incomplete remodeling in the hybrid group, which may contribute to ACL laxity and greater side-to-side differences on KT-1000 testing [29]. In terms of patient-reported outcomes, we assessed IKDC, Lysholm, and Tegner scores, which also showed no significant differences between the groups. These scores are commonly used to measure functional recovery and subjective knee stability, but neither graft type demonstrated a statistically significant advantage. Variability in individual recovery, study design, and follow-up length may have influenced these findings. Overall, while no significant differences were found in patient-reported outcomes or knee stability, further high-quality research is needed to clarify the long-term effectiveness of autografts versus hybrid grafts in ACL reconstruction. When performing ACLR with autologous HTs, small diameter grafts are inevitably encountered due to the variability of tendon size in individuals. HGs that combine undersized autografts with allografts to augment graft diameter are becoming popular, but their clinical outcomes are still controversial. Suboptimal clinical outcomes associated with HGs are not uncommon [18, 26, 27]. And the results are mainly attributed to 3 points based on the review of the included studies. First, slower and less efficient maturation of allografts. Incomplete or delayed revascularization of HGs is detrimental to cell repopulation and tissue reorganization. This difference may lead to knee laxity and reduced mechanical function compared to ACL reconstruction with autografts [59]. Second, inferior tendon-bone healing ability of HGs. Even 2 years after surgery, MRI showed allografts have inferior remodeling in the bone tunnel compared to the autografts [60, 61]. Third, sterilization of grafts with irradiation. Irradiation has a negative effect on the biomechanical properties of the grafts and this adverse effect is dose-dependent [37, 62]. In addition, people with small hamstrings have other risk factors that predispose them to ACL tears, not solely the graft size. For instance, younger age is a risk factor for susceptible ACL tears and is also associated with small diameter grafts. Other factors, such as the relative position of the allograft and autograft in the HG, additional knee injuries and notch impingement, may also affect the final function of the HG [27, 29, 63]. Considering the uncertainty about the clinical risks of HGs, surgeons should fully understand the indications for their use and the amount of allograft supplementation required for undersized HTs. Two of the 14 included studies [22, 26] have recommended that graft diameter for ACLR should be more than 7.5 mm to achieve desired results, and 7 studies [18, 19, 21, 23, 29, 30, 33] defined the cut-off diameter as 8 mm. Larger diameter graft thresholds should be considered for adolescent patients and patients engaged in strenuous activities [64, 65]. Beck et al. identified a simple and easily applied rule via the use of mathematical and cadaveric models [66]. By geometric measurement rather than simple addition, it allows the surgeons to determine the additional size of allograft diameter needed to supplement undersized hamstring autograft. To some extent, it enables surgeons to finalize a more accurate HG diameter. However, many contradictory results have made it unclear whether and when to apply HGs for ACLR to maximize patient benefit. This may limit the further utilization of HGs. Our results did not demonstrate significant differences in any outcomes between the 2 procedures. Nevertheless, HGs are just one of the solutions for small diameter grafts. Most (10/14) included studies expressed concerns about HGs and did not recommend them as a first choice and proposed alternative solutions. Four studies [20, 21, 26, 29] suggested harvesting the bilateral HT; 2 studies [22, 27] suggested tripling the semitendinosus to obtain a wider graft; 4 studies [18, 30, 31, 33] suggested that nonirradiated fresh-frozen grafts may be a better choice if HGs are still employed. The same attention needs to be paid to the challenge of avoiding small diameter grafts intraoperatively. Multiple studies have reported that preoperative anthropometric measurements and imaging can effectively predict intraoperative hamstring graft size [67–70]. Cross-sectional area (CSA) of the ST and GT on MRI and US could reliably predict graft size and assist in surgical planning. Height and weight have frequently been reported to correlate strongly with the size of the HT [71–74]. The combination of CSA and weight predicts up to 29% of the variance [75]. Although each surgeon may have different criteria for small diameter grafts (7 mm, 7.5 mm, or 8 mm), if potential small size grafts are recognized preoperatively, they can choose the appropriate surgical approach based on their experience and patient preference. If a HG was scheduled, the surgeon could only harvest the semitendinosus muscle, rather than both the semitendinosus and gracilis. This preoperatively planned HG reconstruction is beneficial in reducing failure rates [76]. In addition, the quadriceps tendon and peroneus longus tendon, as an alternative to traditional autografts, can provide clinical results comparable to those of autologous HTs [77–80]. The graft diameter can also be augmented by changing the graft folding method [81–83]. Calvo et al. reported a semitendinosus tripling technique that converted a standard quadrupled hamstring graft into a 5-strand graft [82]. At a minimum follow-up of 2 years, the results suggested that the 5-strand hamstring autograft in ACLR was clinically comparable with the quadruple autograft larger than 8 mm. However, it presupposes sufficient graft length (mostly reported as at least 21 cm). Additional studies are needed to assess whether these methods are viable solutions for small diameter grafts. Several reviews have compared two grafts for ACLR. But some included limited studies and sample sizes [24, 84]; some did not provide comprehensive outcomes analyses [25, 85]; some drew improper inclusions [86]. With more strict inclusion criteria and more recent studies [21, 23, 30], we performed a comprehensive review of outcomes, including failure rates, patient-reported outcomes and knee stability. These factors allow us to draw more robust conclusions. But our studies still have several limitations. First, all studies included in this review were level of evidence 2 (1 study) or 3 (13 studies) with a high risk of bias. 10 of the 14 studies included in this review had only fair methodologic quality assessed by the MCMS. Second, the baseline patient characteristics varied across studies, including age, source of allograft, surgical technique, and fixation method, which may influence the outcomes of ACLR. Third, despite outcomes in the included studies were comprehensively reviewed, long-term outcomes and return to sports rates were not found. Fourth, the mean diameter of HTs varied from 7.1 mm to 8.57 mm among the different studies, which may affect the results. One limitation of this study is that it did not compare muscle strength between groups, which could be an important factor in evaluating functional outcomes after ACL reconstruction. Additionally, the duration of surgery was not assessed, but it is worth noting that hybrid graft (HG) procedures may require less operative time compared to autograft procedures. This is due to the fact that harvesting autologous tissue often requires more time, while hybrid grafts can streamline the process. Future studies should consider including these variables to provide a more comprehensive evaluation of outcomes related to muscle strength and surgical efficiency. Conclusion This review found no significant differences in failure rates, knee stability, or patient-reported outcomes between autologous HT and HG in primary ACLR. While HG is a viable alternative for undersized HT grafts, the potential risks of allografts such as immune reactions, delayed healing, and long-term functional failure warrant careful consideration. Surgeons should prioritize obtaining sufficiently sized autografts through optimized evaluations and techniques (e.g., semitendinosus quadrupling) and consider hybrid grafts only as a last-resort option following a comprehensive risk-benefit assessment.